Reduced Matrix Multiplication: Input-Adaptive Matrix-Product Reduction for LLM Inference
Transformer-based language models achieve strong performance but incur substantial inference cost due to repeated high-dimensional matrix multiplications. We propose Reduced Matrix Multiplication (RMM), a training-free, input-adaptive inference method that reduces Transformer matrix products by selecting informative slices along their contraction dimensions, without modifying model weights. Under a simple retention-ratio control, RMM provides a smooth and predictable accuracy-efficiency trade-off. Across language models ranging from 1B to 70B parameters, we find that reduction tolerance depend
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 57%Transformer →
- PossiblePossibly related (embedding) · 53%I shrank a transformer until every number fitted on the screen and made the weights editable [R] →
- FuzzySimilar title/name (fuzzy) · 84%xorbitsai/inference →
“Fuzzy title match (0.92): “Reduced Matrix Multiplication: Input-Adaptive Matrix-Product” ≈ “xorbitsai/inference””
- FuzzyOverlapping authors or contributors · 62%sgl-project/sglang →
“Shared author/contributor keys: zhou”
- LinkedLinked via arxiv author · 85%Zixuan Lan →
“Reduced Matrix Multiplication: Input-Adaptive Matrix-Product Reduction for LLM Inference”
- LinkedLinked via arxiv author · 85%Yanhong Li →
“Reduced Matrix Multiplication: Input-Adaptive Matrix-Product Reduction for LLM Inference”
- LinkedLinked via arxiv author · 85%Jiawei Zhou →
“Reduced Matrix Multiplication: Input-Adaptive Matrix-Product Reduction for LLM Inference”